CLINICAL EVALUATION OF A FINGEROSCILLOMETRIC BLOOD PRESSURE DEVICE Jefferson M. Sesler, Wendy P. Munroe, and James M. McKenney
ABSTRACT: An oscillometric blood pressure (BP) device designed for recording systemic BP from a fmger was compared with a random-zero, mercury-column sphygmomanometer in 76 subjects recruited from the patient population and staff of a primary-care center. After both devices were placed on the left arm or finger of the left hand, three BP readings were obtained with each device in alternating fashion, thereby according random assignment. The mean for all measurements was 119.2n2.4 mm Hg for the finger device and 118.6n5.3 mrn Hg for the randomzero sphygmomanometer; systolic readings were not different (p=0.53) but diastolic readings were different (p=O.OOI). There were considerable differences among the readings obtained with the two devices in individual patients: the mean differences ± tolerance limits for readings obtained with the finger device compared with the random-zero device were 0.6 ± 33.5 mrn Hg for systolic and -2.9 ± 25.3 mrn Hg for diastolic readings. There was also considerable variability in the readings obtained with the fmger device; the minimum to maximum differences in 39 systolic (51 percent) and 26 diastolic readings (34 percent) were greater than 10 mm Hg with the fmger device; 24 systolic (32 percent) and II diastolic readings (14 percent) differed by this amount with the random-zero device. The variability in measurements made with the fmger device was significantly different from the random-zero device for diastolic pressures (p=0.0048) but not for systolic pressures (p=0.8729). Based on the experience obtained in this study, devices to measure BP from the finger cannot be recommended for routine use.
D1ep Ann Pharmacotherl99I;25:1310-4.
IT IS OF CRITICAL IMPORTANCE to have an accurate method of blood pressure (BP) measurement when screening, evaluating, and treating the estimated 58 million Americans who have high Bp'I Accuracy is accomplished by the proper use and maintenance of BP equipment but it also may be affected by the type of equipment used.>' Healthcare workers should be aware of the development and use of new equipment in the evaluation of hypertensive patients. Many patients choose to monitor their own BP. Current estimates are that four to seven percent of all households have a BP device (totaling more than ten million units).5,6 Some devices rely on the patient to palpate and auscultate the Korotkoff sounds; newer equipment measures BP automatically via oscillation (electronic measurement of the JEFFERSON M. SESLER, B.S.Phann., at the time of this research, was a General Resident, Department of Pharmacy Services, Medical College of Virginia Hospitals; he is now a Decentralized Medicine Pharmacist, University of Virginia Hospitals, Charlottesville, VA; WENDY P. MUNROE, Phann.D., is an Assistant Clinical Professor; and JAMES M. MCKENNEY, Phann.D., is a Professor, School of Pharmacy, Medical College of Virginia, Virginia Commonwealth University, P.O. Box 533, Richmond, VA 23298. Reprints: James M. McKenney, Pharm.D.
1310
•
D1ep, The Annals ofPharmacotherapy
•
rise and fall of the BP wave). Self-monitoring may offer several distinct advantages, including improved medication compliance, better BP control, and fewer visits to physicians for monitoring." Self-monitoring also provides health professionals with another means of assessing the effectiveness of medication and determining the proper treatment program.' The introduction of new electronic devices to measure BP from a patient's finger using oscillometric methods could extend the convenience and availability of self-monitoring of BP.IO These instruments are easy to use, compact, and durable. However, a number of factors may affect their accuracy, including fmger temperature, position, and size, as well as peripheral circulation. II The purpose of this study was to compare the accuracy of BP measurements made with a device that records BP using oscillometric methods from a patient's finger with measurements obtained with a mercury-column sphygmomanometer. Methods Adult subjects (aged >20 y) were recruited from patients and staff in a hospital-based, primary-care center. Subjects were excluded if they required greater than a large adult size BP cuff or if they had had a left-side radical mastectomy. The study used the automatic finger oscillometric sphygmomanometers marketed by the PiPeer (model 254) and amron (model HEM812F) companies. Both devices were identical in appearance and were purchased in commercial establishments, Both were battery-powered and had a plastic ring containing the cuff bladder into which a finger is inserted. The bladder was inflated automatically by pressing a button; the pulse rate and the systolic and diastolic BPs were displayed on a digital screen (Figure I). These devices were operated in accordance with the operators' manuals throughout the study. The Hawksley random-zero mercury sphygmomanometer (Hawksley and Sons Limited, Lancing, England) has been previously described in detail." It is a mercury-column manometer that employs auscultation to record Korotkoff sounds. In order to blind the observer to readings and thereby reduce observer bias, it has been modified to contain a reservoir in which variable amounts of mercury are held during each measurement. Thus the "zero" level of the mercury column changes with each measurement and must be read and subtracted from the systolic and diastolic readings after the measurement has been made." Subjects meeting entrance criteria were randomized in block fashion to one of two test sequences. In the first sequence, measurements were taken with the finger sphygmomanometer first and the random-zero mercury sphygmomanometer second; measurements were continued in alternating fashion until a total of three readings had been taken with each device. 14 The second test sequence began measurement with the
1991 December, Volume 25
Research/Practice random zero and alternated with the finger device until a total of three measurements had been taken with each device. All subjects had their BP measured while sining, having been seated for at least five minutes in a quiet room at a comfortable temperature. Each subject's left arm was bared, unrestricted by clothing, with the palm of the hand exposed upward, resting at heart level and supported by an adjacent table. All measurements were taken by trained and experienced investigators who followed the techniques recommended for BP measurement by the American Heart Association." The appropriately sized cuff of the random-zero sphygmomanometer was positioned on the left arm approximately 2.5 ern above the antecubital fossa; the ring cuff of the finger device was positioned around the first digit of the left hand. The brachial artery location was determined by palpation. Prior to each measurement, the random-zero sphygmomanometer was adjusted by spinning the thumbwheel to vary the level of mercury in the reservoir in a random fashion. According to the assigned test sequence, the cuff of the designated device was inflated, the pressures recorded, and the cuff deflated completely for at least 15 seconds before the cuff of the next device was inflated and pressures were recorded. Systolic and diastolic measures were obtained on each patient with the random-zero sphygmomanometer by recording the first and fifth phase of Korotkoff sounds and the level of the mercury column after the recording. This later figure was subtracted from the BP measurement after all study measurements had been obtained. Cuffs of both devices remained in place until all measurements had been completed. The means and standard deviations of the BP readings obtained with each device were determined for each subject, for subgroups, for each measurement, and for the entire study population. Mean BPs for each subject obtained with each device were compared using ANOVA and Tukey's significant difference test. Alpha was set at 0.05. All data were found to be normally distributed prior to using these tests. Differences
between the readings obtained with the finger and the random-zero devices and between minimum and maximum readings obtained with each device for systolic and diastolic BP in each subject were determined. Tolerance limits for the differences between the devices were determined.
Figure I. Schematic representation of a finger blood pressure device.
mm Hg Difference in SBP by Finger Device
40
Table I. Characteristics of the Study Subjects
0 0 0 o 0 o 0 ocP o~o 0
20 0 0
CHARACTERISTIC
NUMBER
0
Sex male female Race white black other Mean age (range) Mean weight Medical problems none hypertension diabetes other
u
td
0
0
0
0
0
On
oo%°el 0
-20 0
oCb
0
0
Do DO
-60
0
o 0
0
0
-40
0
LLl..LL.L.LU--l..LL.L.LU--l...LJ...w....L.Ll-L.L.L.LJL.Ll-L.L.L.LJL.Ll-L.L.L.LJL. L.L.LL.L.LJ
eo
43 (55.6%) 22 (28.9%) 13 (3.9%) 21 (27.6%)
0
0
0
08
0
0
0
0
29 (38.2%) 45 (59.2%) 2(2.6%) 43.3 (21-86 y) 162.8 lb.
0
0
B~ 0
27 (35.5%) 49 (64.5%)
o
0 DO
90
100
110
120
130
140
150
160
170
180
Random Zero Systolic BP (mm Hg) mm Hg Difference in DBP by Finger Device
40 , - - - - - - - - - ' - - - - - - - ' ' - - - - - - - - - - - - - ,
o 30
Table 2. Mean Systolic and Diastolic Blood Pressure Readings (± SD) with the Random-Zero and Digital Devices
20 10
RANDOM-ZERO
Systolic blood pressure measurement I measurement 2 measurement 3 all measurements Diastolic blood pressure measurement I measurement 2 measurement 3 all measurements
'p=O.OOI.
0 0
0
75.7 75.4 74.9 75.3
(12.5) (13.0) (13.7) (12.7)
71.9 72.5 72.9 72.4
(15.2) (13.7) (13.5) (12.7)"
0
DO 0
0
oo§
nBB
°o~ EI ~oB
0 0
-10
0
000
0
0
0
0
EI
60
0 0
Del DO
0
50
0
8~
0
B
-20
0
0 OW
0
0
118.3 (19.3) 119.3 (19.3) 119.2 (18.3) 119.2(18.1)
0
0
DIGITAL
0
119.2 (19.5) 118.3 (18.7) 118.4 (19.0) 118.6 (18.5)
0
0
0
0
70
eo
90
100
110
Random Zero DiastolicBP (mm Hg) Figure 2. Scatter plot of the differences in systolic and diastolic blood pressures recorded with the finger device compared with the random-zero device in each of the 76 study subjects.
D/CP, The Annals ofPharmacotherapy
•
/99/ December, Volume 25 •
/311
The equality of variances (i.e., reproducibility) between the two BP devices was evaluated using a test by Mudholkar and Subbaih (unpublished technical report, University of Rochester and Oakland University).
Results
Seventy-six subjects met entrance criteria and completed all phases of the study protocol (Table I). The majority of subjects were female, black, and without medical problems. Thirty-three subjects had one or more medical problems; 22 had a diagnosis of hypertension and were receiving antihypertensive drug therapy. Approximately 25 percent of the subjects were above their ideal body weight. However, only 6 required the use of a large BP cuff. The mean BP readings with each device are displayed in Table 2. The 119.2 mm Hg mean systolic BP obtained with the fmger device did not differ from the 118.6 mm Hg mean systolic pressure obtained with the random-zero sphygmomanometer (p=0.53). The 72.4 mm Hg mean diastolic BP obtained with the fmger device was statistically different than the 75.3 mm Hg mean diastolic reading obtained with the random zero (p=O.OOI).There was no order or period effect. The mean BP readings obtained with the two devices in subgroups of study subjects are presented in Table 3. Subjects with medical problems, many of whom were hypertensive, tended to have higher readings with both devices than did subjects without medical problems. Mean readings obtained with the random-zero device were generally higher than those obtained with the fmger device. Figure 2 indicates considerable differences between mean systolic and diastolic BPs recorded with the finger device and those recorded with the random-zero devices. The average differences ± standard deviation between these readings were 0.6 ± 14.2 mm Hg for systolic and -2.9 ± 10.7 for diastolic pressures. The tolerance limit of ± 33.5 mm Hg for systolic and ± 25.3 mm Hg for diastolic pressures indicates that a systolic BP obtained with the finger device would be between -32.9 and +34.1 mm Hg of the reading obtained with a random-zero device 95 percent of the time; a diastolic BP would be between -28.2 and +22.4 mm Hg of the results recorded with the random-zero device 95 percent of the time. BP differences greater than 10 mm Hg between the two devices were common; some were as great as 30 mm Hg. The distribution of these differences is presented in Figure 3. Forty-six percent of systolic readings and 34 percent of diastolic readings with the finger device were greater than
± 10 mm Hg of the readings obtained with the random-zero manometer. Differences between the minimum and maximum systolic and diastolic BPs measured with each device in each subject are presented in Figure 4. Thirty-nine systolic (5 I percent) and 26 diastolic readings (34 percent) differed by more than 10 mm Hg when the finger device was used. This same disparity occurred with 24 systolic (32 percent) and I I diastolic readings (14 percent) with the randomzero device. Using the likelihood ratio statistic of Mudholkar and Subbaih, the variability in diastolic BPs obtained with the fmger device was found to be significantly different from that obtained with the random-zero device (p=0.0048); no differences were found in the variability of systolic BPs between the two devices (p=0.8729). Discussion
The fmger devices tested in this study fail to provide BP readings that are similar to those obtained with the random-zero, mercury-column manometer. Although statistical difference was found only between diastolic readings for the study group, there were considerable differences in both the systolic and diastolic readings between the two devices in individual patients. The variability of diastolic BP readings was significantly greater with the finger device than with that obtained with the random-zero device.
Subjects (n)
16~--=-----~--------------------'
~
~
4
~
~
~
~
~
~
Systolic BP Differences (mm Hg) Subjects (n) 16~-------------------,
14 12
Table 3. Mean Systolic and Diastolic Blood Pressures in Subgroups of the Study Population Recorded with the Random-Zero and Digital Devices
10
8 6
RANDOM-ZERO CATEGORIES
Medical problems present absent Diagnosis of hypertension present absent
n
S
D
43 33
131.4 108.7
80.3 71.5
DIGITAL S
4 2
125.3 75.1 114.5 70.4
0
'-20
-20
-10
~
+1
+15
+10
+20
>+20
Diastolic BP Differences (mm Hg)
22 54
134.1 112.3
81.8 72.7
126.7 76.2 116.1 70.9
D = diastolic; S = systolic.
1312 •
D
Dlt.P, The Annals ofPharmacotherapy
•
Figure 3. Bar graph representing the distributions of differences between systolic and diastolic blood pressures recorded with the finger blood pressure device compared with the random-zero device.
1991 December. Volume 25
Research/Practice
These fmdings are in substantial agreement with most,I6-18 but not all," of the published literature regarding BP measurements made from a fmger. Although the fmger devices used in these studies had different design features, all used the oscillometric method of obtaining BP readings. A very early version of these devices was evaluated in 75 patients by Weaver and Bohr who found it to underestimate by 5 to 31 mm Hg the BP obtained by auscultation." More recently, Watkins-Pitchford reported a mean BP of I 32n9 mm Hg with a mercury auscultation method and 152/89 mm Hg with the Healthcheck fmger BP device in 50 patients in a recovery room." It has been suggested that differences in BP among the devices may not be a result of differences in the devices themselves, but rather of differences between the oscillation and auscultation methods of obtaining BP.20.21 However, Jenner et al. refuted this theory when they used the oscillation method to obtain BP readings from the brachial artery and a finger in 119 normotensive and 76 hypertensive patients and still found that the fmger device substantially underestimated the pressures obtained from the brachial artery." Another explanation given for differences in pressures between devices that use the fmger and the arm may be inherent differences between brachial and peripheral pres-
sures. Nielsen et aI., for example, reported that the mean intraarterial BPs recorded in 13 normal subjects were 128nl mm Hg when measurements were taken from the brachial artery and 154n2 mm Hg when measurements were taken from the posterior tibial artery." Variables known to affect the accuracy of BP measurements made with a finger device include poor peripheral circulation, fmger temperature (cold fmgers will yield low readings), and position (index fmger should be at heart level)." We attempted to minimize these variables by maintaining the subject's hand at heart level and providing a warm study environment. It is quite likely that finger BP devices will be very attractive to the public. They are easy to use and generate clear readings in minutes. However, based on this study, these devices cannot be recommended. In fact, if patients use them to monitor their BP and to reinforce the need for adherence with prescribed medication regimens, they are likely to make significant misinterpretations. Potentially dangerous decisions can occur if patients report results of pressures measured with fmger devices to physicians who use them to alter treatment regimens. ~ References I. 1988 Joint National Committee. The 1988 report of the Joint National
2.
Subjects (n) 40,------------,----------,
3.
30
4.
20
5. 6.
Committee on detection, evaluation and treatment of high blood pressure. Arch InternMed 1988;148:1023-38. HUNT JC, FROHLICH ED, MOSER M, et al. Devices used for self-measurement of blood pressure: revised statement of the National High Blood Pressure Education Program. Arch InternMed 1985;145:2231-4. Health and Public Policy Committee. Automated ambulatory blood pressure monitoring. Ann InternMed 1986;104:275-8. Association for the Advancement of Medical Instrumentation. American national stand for nonautomated sphygmomanometers. Arlington, VA: Association for the Advancement of Medical Instrumentation, 1986. Two of every ten households now use home diagnostic tests. Am Pharm I987;NS27: 15. HAHN LP, FOLSOM AR, SPRAFKA JM, PRINEAS RJ. Prevalence and accuracy of home sphygmomanometers in an urban population. Am 1
PublicHealth 1987;77:1459-61.
10
7. HAYNESRB, SACKETDL, GIBSON ES, et al. Improvement of medication compliance in uncontrolled hypertension. Lancet 1976;/:1265-8.
0..-----20
16-20 11-15 5-10
'5
>20 16-20 11-15 5-10
'5
Systolic BP Differences(mm Hg) Subjects (n)
35,------------,-------------,
30
8. EDMONDS0, FOERSTERE, GROTH H, GREMINGERP, SIEGENTHALER W, VETTER W. Does self-measurement of blood pressure improve pa-
tient compliance in hypertension? 1 Hypertens 1985;3(suppl 1):31-4. 9. BURNS-COX CJ, REES JR, WILSON RSE. Pilot study of home measurement of blood pressure by hypertensive patients. Br Med 1 1975;3:80. 10. SCHMIDTGR, WENIG JH. An evaluation of home blood-pressure monitoring devices. Am Pharm 1989;NS29:25-30. II. WESSLINGKH, SETTELS JJ, VAN DER HOEVENGM, et al. Effects of peripheral vasoconstriction on the measurement of blood pressure in a
finger, Cardiovasc Res 1985;19: 139-45. 25
>20 16-20 11-15 6-10
'6
>20 16-20 11-16 5-10
'5
Diastolic BP Differences(mm Hg) Random Zero BPs on Left; Finger Device BPs on Right Figure 4. Bar graph of the differences between the minimum and maximum systolic and diastolic blood pressures recorded in each of the 76 study subjects with the finger and random-zero blood pressure devices.
12. WRIGHT BM. DORE CF. A random-zero sphygmomanometer. Lancet 1970;1 :337-8. 13. PARKER0, LIU K, DYER AR, GIUMETTI0, LIAO Y, STAMLERJ. A comparison of the random-zero and standard mercury sphygmomanometers. Hypertension 1988;11:269-72. 14. LLABRE MM, IRONSONGH, SPIlZER SB, GELLMAN MD, WEIDLEROJ. SCHNEIDERMAN N. How many blood pressure measurements are enough? An application of generalizability theory to the study of blood pressure reliability. Psychophysiology 1988;25:97-105. 15. KIRKENDALL WM, FEINLEIBM, FRIES ED, et al. American Heart Association recommendations for human blood pressure determination by sphygmomanometer. Circulation 1980;62:I I46A-55A. 16. WEAVER JC, BOHR OF. The digital blood pressure. 1. Values in normal subjects. Am Heart1 1950;39:413-22. 17. WATKINS-PITCHFORD JM. The performance of a digital blood pressure monitoring device. lAMA 1989;261:1153-4. 18. MOLHOEK PG. WESSELING KH, ARNlZENIUS AC, et al. Initial results of
D/CP, The Annals ofPharmacotherapy
•
/99/ December, Volume 25 •
/313
non-invasive measurement of finger blood pressure according to Penaz. Automedica 1983;4:241-6. 19. IMAI Y. NIHEI M, ABE K. et aI. A fingervolume-oscillometric device for
monitoring ambulatory bloodpressure: laboratory andclinical evaluations. Clin Exp Hypertens [A] 1987;A9:2001-25. 20. NAKAYMA R. AZUMA T. Noninvasive measurements of digitalarterial pressure andcompliance in man. Am J PhysioI1977;233:HI68-78. 21. HARRISON DW, KELLYPL. Homehealth care:accuracy, calibration, exhaust,and failure ratecomparisons of digital blood pressure monitors. Med Instrumentation 1987;21:323-8.
22. JENNER DA, BEILIN LJ, VANDONGENR. A comparison of bloodpressure measurements obtained with the dinamap 845XT, the standard mercury sphygmomanometer and the London Schoolof Hygiene device. Clin Exp Hypertens [A] 1988;AIO:575-88. 23. NIELSEN PE, BARRASJP, GOLSTEIN P. Systolic pressure amplification in the arteries of normal subjects. Scand J Clin Lab Invest 1974;33:371-7.
con el esfigmoman6metrose encontr6 una diferenciapromedio ± el limitede toleranciade 0.6 ± 33.5 mm Hg para la presi6n sist61ica y -2.9 ± 25.3 mm Hg para la presi6n diast6lica. Se encontrarontambien variacionesconsiderablesentre las lecturasde un mismo paciente cuando se utilizabael oscil6metrodigital. En 39 (5I por ciento) de las lecturassist61icas y en 26 (34 por ciento) de las lecturasdiast61icas hubo mas de 10 mm Hg de diferenciaentre las lecturashechas por el instrumentodigital. Utilizandoel esfigmoman6metro, diferencias mayores a IO mm Hg s610 ocurrieronen 24 (32 por ciento) de las lecturassist61icas y I I (14 por ciento) de las lecturasdiast6licas. Estas variacionesen las lecturasobtenidas utilizandoel instrumentodigital resultaronser significativamente diferentesa las encontradascon el esfigmoman6metropara las presionesdiast61icas (p=0.0048), pero no en las sist61icas (p=0.8729). Concluyen los autores que, basado en estos hallazgos, los instrumentospara medir la presi6n sanguineapor del dedo no se pueden recomendarpara uso rutinario. MIRZA D. MARTiNEZ
RESUME
EXTRACTO
La precisi6ndel rnetodoutilizadoal determinarla presi6nsanguineade un individuoes esencial para poder detectar,evaluar y tratar adecuadarnente a personascon presi6n sanguineaalta. La precisi6nde la lecturade la presi6n dependedel equipo utilizado, el procedimiento, y el mantenimiento que se provea aI instrumento. Algunosde los instrumentos utilizadospara medir la presi6nsanguineadependen de que el individuopueda palpar (sentir)y auscultar (ofr) los sonidos Korotkoffmientrasque otros instrumentos miden la presi6n sanguinea automaticamente medianteoscilometria,esto es, determinando electr6nicamente el alza y baja de la onda de presi6n sanguinea.Estos nuevos instrumentos electr6nicosque miden la presi6n mediante oscilometriapueden representarpara muchos pacientes una forma convenientey accesiblede proveer seguimientoa su presi6n sanguinea. Estos autores compararon la precisi6nde las lecturasde presi6n sanguineahechas utilizandoun instrumentodigital que detennina la presi6n sanguineadel dedo por metodos oscilometricoscon las lecturas obtenidaspor un esfigmoman6metro. En el estudio se reclutaron76 sujetos,pacientes y personalde un centro de cuidado primario.Se colocaronambos instrumentosde medir presi6n en el brazo Izquierdo y en el dedo de la mano izquierday se tomaron tres lecturasde presi6n sanguineacon cada instrumentoen forma a1tema. EI promedio de todas las lecturasutilizandoel instrumentodigital fue I 19.2/72.4mientrasque con el esfigmoman6metro fue de 118.6/75.3. Las lecturasde presi6n sist61ica no fueron diferentesentre los dos instrumentos(p=0.53), pero las lecturasde presi6n diast61ica resultaronser significativamente diferentes(p=O.OO I). Tambien existierondiferenciasconsiderablesentre las lecturasde presi6n obtenidasen un mismo paciente cuando se usaban los diferentesinstrumentos. AI comparar el oscil6metrodigital
1314 •
Dlep, The Annals ofPharmacotherapy
•
Un appareilde mesure de la tension arteriellepar oscillometrie, dessine pour evaluer la tension arteriellesystemiqueau doigt, a ete compare 11 un sphygmomanometre (SM) 11 colonne de mercure chez 76 sujets patients ou employesd'une c1inique de soins. Apres que les deux appareils eurent ete mis en place, soit sur Ie bras gauche, soit sur un doigt de la main gauche, trois mesures de la tension arterielleont ete obtenues avec chaque appareilen a1temance selon une sequence pre-randomisee. La moyenne de toutes les mesuresetait de I 19.2(72.4 mm Hg pour l'appareil au doigt et I 18.6(75.3 mm Hg pour Ie SM; les lecturesde la tensionsystolique(TS) etaient Iressemblables(p=0.53) mais les lecturesde tension diastolique(TO) etaient differentes(p=O.OOI). On a note des ecarts appreciables entre les lecturesobtenues par les deux appareilschez un merne sujet; les differencesmoyennes (± limitesde tolerance)pour les lecturesobtenues avec l'appareil au doigt comparees au SM etaient de 0.6 ± 33.5 mm Hg pour la TS et -2.9 ± 25.3 mm Hg pour la TO. On a observe aussi une grande variabilite dans les lectures obtenues avec l'appareil au doigt: les differencesentre les minima et maxima dans 39 lecturesde TS (5I pourcent)et 26 lecturesde TO (34 pourcent)etaient de plus de IOmm Hg avec l'appareil au doigt, tandis que 24 lecturesde TS (32 pourcent)et I I lecturesde TO (14 pourcent) differaientelles aussi de plus de IOmm Hg avec Ie SM. En conclusion, la variabilitedans les mesures faites avec l'appareil au doigt etaient significativement differentesde celles prises au sphygmornanometre pour les tensionsdiastoliques(p=0.0048) mais non pour les tensions systoliques(p=0.8729). En se basant sur les resultatsde cette etude, les auteurs ne peuvent recommanderde routine I'emploi d' appareilde mesure de la tension arterielleau doigt.
/991 December, Volume 25
DENYSE DEMERS
ABSTRACT: An oscillometric blood pressure (BP) device designed for recording systemic BP from a fmger was compared with a random-zero, mercury-column sphygmomanometer in 76 subjects recruited from the patient population and staff of a primary-care center. After both devices were placed on the left arm or finger of the left hand, three BP readings were obtained with each device in alternating fashion, thereby according random assignment. The mean for all measurements was 119.2n2.4 mm Hg for the finger device and 118.6n5.3 mrn Hg for the randomzero sphygmomanometer; systolic readings were not different (p=0.53) but diastolic readings were different (p=O.OOI). There were considerable differences among the readings obtained with the two devices in individual patients: the mean differences ± tolerance limits for readings obtained with the finger device compared with the random-zero device were 0.6 ± 33.5 mrn Hg for systolic and -2.9 ± 25.3 mrn Hg for diastolic readings. There was also considerable variability in the readings obtained with the fmger device; the minimum to maximum differences in 39 systolic (51 percent) and 26 diastolic readings (34 percent) were greater than 10 mm Hg with the fmger device; 24 systolic (32 percent) and II diastolic readings (14 percent) differed by this amount with the random-zero device. The variability in measurements made with the fmger device was significantly different from the random-zero device for diastolic pressures (p=0.0048) but not for systolic pressures (p=0.8729). Based on the experience obtained in this study, devices to measure BP from the finger cannot be recommended for routine use.
D1ep Ann Pharmacotherl99I;25:1310-4.
IT IS OF CRITICAL IMPORTANCE to have an accurate method of blood pressure (BP) measurement when screening, evaluating, and treating the estimated 58 million Americans who have high Bp'I Accuracy is accomplished by the proper use and maintenance of BP equipment but it also may be affected by the type of equipment used.>' Healthcare workers should be aware of the development and use of new equipment in the evaluation of hypertensive patients. Many patients choose to monitor their own BP. Current estimates are that four to seven percent of all households have a BP device (totaling more than ten million units).5,6 Some devices rely on the patient to palpate and auscultate the Korotkoff sounds; newer equipment measures BP automatically via oscillation (electronic measurement of the JEFFERSON M. SESLER, B.S.Phann., at the time of this research, was a General Resident, Department of Pharmacy Services, Medical College of Virginia Hospitals; he is now a Decentralized Medicine Pharmacist, University of Virginia Hospitals, Charlottesville, VA; WENDY P. MUNROE, Phann.D., is an Assistant Clinical Professor; and JAMES M. MCKENNEY, Phann.D., is a Professor, School of Pharmacy, Medical College of Virginia, Virginia Commonwealth University, P.O. Box 533, Richmond, VA 23298. Reprints: James M. McKenney, Pharm.D.
1310
•
D1ep, The Annals ofPharmacotherapy
•
rise and fall of the BP wave). Self-monitoring may offer several distinct advantages, including improved medication compliance, better BP control, and fewer visits to physicians for monitoring." Self-monitoring also provides health professionals with another means of assessing the effectiveness of medication and determining the proper treatment program.' The introduction of new electronic devices to measure BP from a patient's finger using oscillometric methods could extend the convenience and availability of self-monitoring of BP.IO These instruments are easy to use, compact, and durable. However, a number of factors may affect their accuracy, including fmger temperature, position, and size, as well as peripheral circulation. II The purpose of this study was to compare the accuracy of BP measurements made with a device that records BP using oscillometric methods from a patient's finger with measurements obtained with a mercury-column sphygmomanometer. Methods Adult subjects (aged >20 y) were recruited from patients and staff in a hospital-based, primary-care center. Subjects were excluded if they required greater than a large adult size BP cuff or if they had had a left-side radical mastectomy. The study used the automatic finger oscillometric sphygmomanometers marketed by the PiPeer (model 254) and amron (model HEM812F) companies. Both devices were identical in appearance and were purchased in commercial establishments, Both were battery-powered and had a plastic ring containing the cuff bladder into which a finger is inserted. The bladder was inflated automatically by pressing a button; the pulse rate and the systolic and diastolic BPs were displayed on a digital screen (Figure I). These devices were operated in accordance with the operators' manuals throughout the study. The Hawksley random-zero mercury sphygmomanometer (Hawksley and Sons Limited, Lancing, England) has been previously described in detail." It is a mercury-column manometer that employs auscultation to record Korotkoff sounds. In order to blind the observer to readings and thereby reduce observer bias, it has been modified to contain a reservoir in which variable amounts of mercury are held during each measurement. Thus the "zero" level of the mercury column changes with each measurement and must be read and subtracted from the systolic and diastolic readings after the measurement has been made." Subjects meeting entrance criteria were randomized in block fashion to one of two test sequences. In the first sequence, measurements were taken with the finger sphygmomanometer first and the random-zero mercury sphygmomanometer second; measurements were continued in alternating fashion until a total of three readings had been taken with each device. 14 The second test sequence began measurement with the
1991 December, Volume 25
Research/Practice random zero and alternated with the finger device until a total of three measurements had been taken with each device. All subjects had their BP measured while sining, having been seated for at least five minutes in a quiet room at a comfortable temperature. Each subject's left arm was bared, unrestricted by clothing, with the palm of the hand exposed upward, resting at heart level and supported by an adjacent table. All measurements were taken by trained and experienced investigators who followed the techniques recommended for BP measurement by the American Heart Association." The appropriately sized cuff of the random-zero sphygmomanometer was positioned on the left arm approximately 2.5 ern above the antecubital fossa; the ring cuff of the finger device was positioned around the first digit of the left hand. The brachial artery location was determined by palpation. Prior to each measurement, the random-zero sphygmomanometer was adjusted by spinning the thumbwheel to vary the level of mercury in the reservoir in a random fashion. According to the assigned test sequence, the cuff of the designated device was inflated, the pressures recorded, and the cuff deflated completely for at least 15 seconds before the cuff of the next device was inflated and pressures were recorded. Systolic and diastolic measures were obtained on each patient with the random-zero sphygmomanometer by recording the first and fifth phase of Korotkoff sounds and the level of the mercury column after the recording. This later figure was subtracted from the BP measurement after all study measurements had been obtained. Cuffs of both devices remained in place until all measurements had been completed. The means and standard deviations of the BP readings obtained with each device were determined for each subject, for subgroups, for each measurement, and for the entire study population. Mean BPs for each subject obtained with each device were compared using ANOVA and Tukey's significant difference test. Alpha was set at 0.05. All data were found to be normally distributed prior to using these tests. Differences
between the readings obtained with the finger and the random-zero devices and between minimum and maximum readings obtained with each device for systolic and diastolic BP in each subject were determined. Tolerance limits for the differences between the devices were determined.
Figure I. Schematic representation of a finger blood pressure device.
mm Hg Difference in SBP by Finger Device
40
Table I. Characteristics of the Study Subjects
0 0 0 o 0 o 0 ocP o~o 0
20 0 0
CHARACTERISTIC
NUMBER
0
Sex male female Race white black other Mean age (range) Mean weight Medical problems none hypertension diabetes other
u
td
0
0
0
0
0
On
oo%°el 0
-20 0
oCb
0
0
Do DO
-60
0
o 0
0
0
-40
0
LLl..LL.L.LU--l..LL.L.LU--l...LJ...w....L.Ll-L.L.L.LJL.Ll-L.L.L.LJL.Ll-L.L.L.LJL. L.L.LL.L.LJ
eo
43 (55.6%) 22 (28.9%) 13 (3.9%) 21 (27.6%)
0
0
0
08
0
0
0
0
29 (38.2%) 45 (59.2%) 2(2.6%) 43.3 (21-86 y) 162.8 lb.
0
0
B~ 0
27 (35.5%) 49 (64.5%)
o
0 DO
90
100
110
120
130
140
150
160
170
180
Random Zero Systolic BP (mm Hg) mm Hg Difference in DBP by Finger Device
40 , - - - - - - - - - ' - - - - - - - ' ' - - - - - - - - - - - - - ,
o 30
Table 2. Mean Systolic and Diastolic Blood Pressure Readings (± SD) with the Random-Zero and Digital Devices
20 10
RANDOM-ZERO
Systolic blood pressure measurement I measurement 2 measurement 3 all measurements Diastolic blood pressure measurement I measurement 2 measurement 3 all measurements
'p=O.OOI.
0 0
0
75.7 75.4 74.9 75.3
(12.5) (13.0) (13.7) (12.7)
71.9 72.5 72.9 72.4
(15.2) (13.7) (13.5) (12.7)"
0
DO 0
0
oo§
nBB
°o~ EI ~oB
0 0
-10
0
000
0
0
0
0
EI
60
0 0
Del DO
0
50
0
8~
0
B
-20
0
0 OW
0
0
118.3 (19.3) 119.3 (19.3) 119.2 (18.3) 119.2(18.1)
0
0
DIGITAL
0
119.2 (19.5) 118.3 (18.7) 118.4 (19.0) 118.6 (18.5)
0
0
0
0
70
eo
90
100
110
Random Zero DiastolicBP (mm Hg) Figure 2. Scatter plot of the differences in systolic and diastolic blood pressures recorded with the finger device compared with the random-zero device in each of the 76 study subjects.
D/CP, The Annals ofPharmacotherapy
•
/99/ December, Volume 25 •
/311
The equality of variances (i.e., reproducibility) between the two BP devices was evaluated using a test by Mudholkar and Subbaih (unpublished technical report, University of Rochester and Oakland University).
Results
Seventy-six subjects met entrance criteria and completed all phases of the study protocol (Table I). The majority of subjects were female, black, and without medical problems. Thirty-three subjects had one or more medical problems; 22 had a diagnosis of hypertension and were receiving antihypertensive drug therapy. Approximately 25 percent of the subjects were above their ideal body weight. However, only 6 required the use of a large BP cuff. The mean BP readings with each device are displayed in Table 2. The 119.2 mm Hg mean systolic BP obtained with the fmger device did not differ from the 118.6 mm Hg mean systolic pressure obtained with the random-zero sphygmomanometer (p=0.53). The 72.4 mm Hg mean diastolic BP obtained with the fmger device was statistically different than the 75.3 mm Hg mean diastolic reading obtained with the random zero (p=O.OOI).There was no order or period effect. The mean BP readings obtained with the two devices in subgroups of study subjects are presented in Table 3. Subjects with medical problems, many of whom were hypertensive, tended to have higher readings with both devices than did subjects without medical problems. Mean readings obtained with the random-zero device were generally higher than those obtained with the fmger device. Figure 2 indicates considerable differences between mean systolic and diastolic BPs recorded with the finger device and those recorded with the random-zero devices. The average differences ± standard deviation between these readings were 0.6 ± 14.2 mm Hg for systolic and -2.9 ± 10.7 for diastolic pressures. The tolerance limit of ± 33.5 mm Hg for systolic and ± 25.3 mm Hg for diastolic pressures indicates that a systolic BP obtained with the finger device would be between -32.9 and +34.1 mm Hg of the reading obtained with a random-zero device 95 percent of the time; a diastolic BP would be between -28.2 and +22.4 mm Hg of the results recorded with the random-zero device 95 percent of the time. BP differences greater than 10 mm Hg between the two devices were common; some were as great as 30 mm Hg. The distribution of these differences is presented in Figure 3. Forty-six percent of systolic readings and 34 percent of diastolic readings with the finger device were greater than
± 10 mm Hg of the readings obtained with the random-zero manometer. Differences between the minimum and maximum systolic and diastolic BPs measured with each device in each subject are presented in Figure 4. Thirty-nine systolic (5 I percent) and 26 diastolic readings (34 percent) differed by more than 10 mm Hg when the finger device was used. This same disparity occurred with 24 systolic (32 percent) and I I diastolic readings (14 percent) with the randomzero device. Using the likelihood ratio statistic of Mudholkar and Subbaih, the variability in diastolic BPs obtained with the fmger device was found to be significantly different from that obtained with the random-zero device (p=0.0048); no differences were found in the variability of systolic BPs between the two devices (p=0.8729). Discussion
The fmger devices tested in this study fail to provide BP readings that are similar to those obtained with the random-zero, mercury-column manometer. Although statistical difference was found only between diastolic readings for the study group, there were considerable differences in both the systolic and diastolic readings between the two devices in individual patients. The variability of diastolic BP readings was significantly greater with the finger device than with that obtained with the random-zero device.
Subjects (n)
16~--=-----~--------------------'
~
~
4
~
~
~
~
~
~
Systolic BP Differences (mm Hg) Subjects (n) 16~-------------------,
14 12
Table 3. Mean Systolic and Diastolic Blood Pressures in Subgroups of the Study Population Recorded with the Random-Zero and Digital Devices
10
8 6
RANDOM-ZERO CATEGORIES
Medical problems present absent Diagnosis of hypertension present absent
n
S
D
43 33
131.4 108.7
80.3 71.5
DIGITAL S
4 2
125.3 75.1 114.5 70.4
0
'-20
-20
-10
~
+1
+15
+10
+20
>+20
Diastolic BP Differences (mm Hg)
22 54
134.1 112.3
81.8 72.7
126.7 76.2 116.1 70.9
D = diastolic; S = systolic.
1312 •
D
Dlt.P, The Annals ofPharmacotherapy
•
Figure 3. Bar graph representing the distributions of differences between systolic and diastolic blood pressures recorded with the finger blood pressure device compared with the random-zero device.
1991 December. Volume 25
Research/Practice
These fmdings are in substantial agreement with most,I6-18 but not all," of the published literature regarding BP measurements made from a fmger. Although the fmger devices used in these studies had different design features, all used the oscillometric method of obtaining BP readings. A very early version of these devices was evaluated in 75 patients by Weaver and Bohr who found it to underestimate by 5 to 31 mm Hg the BP obtained by auscultation." More recently, Watkins-Pitchford reported a mean BP of I 32n9 mm Hg with a mercury auscultation method and 152/89 mm Hg with the Healthcheck fmger BP device in 50 patients in a recovery room." It has been suggested that differences in BP among the devices may not be a result of differences in the devices themselves, but rather of differences between the oscillation and auscultation methods of obtaining BP.20.21 However, Jenner et al. refuted this theory when they used the oscillation method to obtain BP readings from the brachial artery and a finger in 119 normotensive and 76 hypertensive patients and still found that the fmger device substantially underestimated the pressures obtained from the brachial artery." Another explanation given for differences in pressures between devices that use the fmger and the arm may be inherent differences between brachial and peripheral pres-
sures. Nielsen et aI., for example, reported that the mean intraarterial BPs recorded in 13 normal subjects were 128nl mm Hg when measurements were taken from the brachial artery and 154n2 mm Hg when measurements were taken from the posterior tibial artery." Variables known to affect the accuracy of BP measurements made with a finger device include poor peripheral circulation, fmger temperature (cold fmgers will yield low readings), and position (index fmger should be at heart level)." We attempted to minimize these variables by maintaining the subject's hand at heart level and providing a warm study environment. It is quite likely that finger BP devices will be very attractive to the public. They are easy to use and generate clear readings in minutes. However, based on this study, these devices cannot be recommended. In fact, if patients use them to monitor their BP and to reinforce the need for adherence with prescribed medication regimens, they are likely to make significant misinterpretations. Potentially dangerous decisions can occur if patients report results of pressures measured with fmger devices to physicians who use them to alter treatment regimens. ~ References I. 1988 Joint National Committee. The 1988 report of the Joint National
2.
Subjects (n) 40,------------,----------,
3.
30
4.
20
5. 6.
Committee on detection, evaluation and treatment of high blood pressure. Arch InternMed 1988;148:1023-38. HUNT JC, FROHLICH ED, MOSER M, et al. Devices used for self-measurement of blood pressure: revised statement of the National High Blood Pressure Education Program. Arch InternMed 1985;145:2231-4. Health and Public Policy Committee. Automated ambulatory blood pressure monitoring. Ann InternMed 1986;104:275-8. Association for the Advancement of Medical Instrumentation. American national stand for nonautomated sphygmomanometers. Arlington, VA: Association for the Advancement of Medical Instrumentation, 1986. Two of every ten households now use home diagnostic tests. Am Pharm I987;NS27: 15. HAHN LP, FOLSOM AR, SPRAFKA JM, PRINEAS RJ. Prevalence and accuracy of home sphygmomanometers in an urban population. Am 1
PublicHealth 1987;77:1459-61.
10
7. HAYNESRB, SACKETDL, GIBSON ES, et al. Improvement of medication compliance in uncontrolled hypertension. Lancet 1976;/:1265-8.
0..-----20
16-20 11-15 5-10
'5
>20 16-20 11-15 5-10
'5
Systolic BP Differences(mm Hg) Subjects (n)
35,------------,-------------,
30
8. EDMONDS0, FOERSTERE, GROTH H, GREMINGERP, SIEGENTHALER W, VETTER W. Does self-measurement of blood pressure improve pa-
tient compliance in hypertension? 1 Hypertens 1985;3(suppl 1):31-4. 9. BURNS-COX CJ, REES JR, WILSON RSE. Pilot study of home measurement of blood pressure by hypertensive patients. Br Med 1 1975;3:80. 10. SCHMIDTGR, WENIG JH. An evaluation of home blood-pressure monitoring devices. Am Pharm 1989;NS29:25-30. II. WESSLINGKH, SETTELS JJ, VAN DER HOEVENGM, et al. Effects of peripheral vasoconstriction on the measurement of blood pressure in a
finger, Cardiovasc Res 1985;19: 139-45. 25
>20 16-20 11-15 6-10
'6
>20 16-20 11-16 5-10
'5
Diastolic BP Differences(mm Hg) Random Zero BPs on Left; Finger Device BPs on Right Figure 4. Bar graph of the differences between the minimum and maximum systolic and diastolic blood pressures recorded in each of the 76 study subjects with the finger and random-zero blood pressure devices.
12. WRIGHT BM. DORE CF. A random-zero sphygmomanometer. Lancet 1970;1 :337-8. 13. PARKER0, LIU K, DYER AR, GIUMETTI0, LIAO Y, STAMLERJ. A comparison of the random-zero and standard mercury sphygmomanometers. Hypertension 1988;11:269-72. 14. LLABRE MM, IRONSONGH, SPIlZER SB, GELLMAN MD, WEIDLEROJ. SCHNEIDERMAN N. How many blood pressure measurements are enough? An application of generalizability theory to the study of blood pressure reliability. Psychophysiology 1988;25:97-105. 15. KIRKENDALL WM, FEINLEIBM, FRIES ED, et al. American Heart Association recommendations for human blood pressure determination by sphygmomanometer. Circulation 1980;62:I I46A-55A. 16. WEAVER JC, BOHR OF. The digital blood pressure. 1. Values in normal subjects. Am Heart1 1950;39:413-22. 17. WATKINS-PITCHFORD JM. The performance of a digital blood pressure monitoring device. lAMA 1989;261:1153-4. 18. MOLHOEK PG. WESSELING KH, ARNlZENIUS AC, et al. Initial results of
D/CP, The Annals ofPharmacotherapy
•
/99/ December, Volume 25 •
/313
non-invasive measurement of finger blood pressure according to Penaz. Automedica 1983;4:241-6. 19. IMAI Y. NIHEI M, ABE K. et aI. A fingervolume-oscillometric device for
monitoring ambulatory bloodpressure: laboratory andclinical evaluations. Clin Exp Hypertens [A] 1987;A9:2001-25. 20. NAKAYMA R. AZUMA T. Noninvasive measurements of digitalarterial pressure andcompliance in man. Am J PhysioI1977;233:HI68-78. 21. HARRISON DW, KELLYPL. Homehealth care:accuracy, calibration, exhaust,and failure ratecomparisons of digital blood pressure monitors. Med Instrumentation 1987;21:323-8.
22. JENNER DA, BEILIN LJ, VANDONGENR. A comparison of bloodpressure measurements obtained with the dinamap 845XT, the standard mercury sphygmomanometer and the London Schoolof Hygiene device. Clin Exp Hypertens [A] 1988;AIO:575-88. 23. NIELSEN PE, BARRASJP, GOLSTEIN P. Systolic pressure amplification in the arteries of normal subjects. Scand J Clin Lab Invest 1974;33:371-7.
con el esfigmoman6metrose encontr6 una diferenciapromedio ± el limitede toleranciade 0.6 ± 33.5 mm Hg para la presi6n sist61ica y -2.9 ± 25.3 mm Hg para la presi6n diast6lica. Se encontrarontambien variacionesconsiderablesentre las lecturasde un mismo paciente cuando se utilizabael oscil6metrodigital. En 39 (5I por ciento) de las lecturassist61icas y en 26 (34 por ciento) de las lecturasdiast61icas hubo mas de 10 mm Hg de diferenciaentre las lecturashechas por el instrumentodigital. Utilizandoel esfigmoman6metro, diferencias mayores a IO mm Hg s610 ocurrieronen 24 (32 por ciento) de las lecturassist61icas y I I (14 por ciento) de las lecturasdiast6licas. Estas variacionesen las lecturasobtenidas utilizandoel instrumentodigital resultaronser significativamente diferentesa las encontradascon el esfigmoman6metropara las presionesdiast61icas (p=0.0048), pero no en las sist61icas (p=0.8729). Concluyen los autores que, basado en estos hallazgos, los instrumentospara medir la presi6n sanguineapor del dedo no se pueden recomendarpara uso rutinario. MIRZA D. MARTiNEZ
RESUME
EXTRACTO
La precisi6ndel rnetodoutilizadoal determinarla presi6nsanguineade un individuoes esencial para poder detectar,evaluar y tratar adecuadarnente a personascon presi6n sanguineaalta. La precisi6nde la lecturade la presi6n dependedel equipo utilizado, el procedimiento, y el mantenimiento que se provea aI instrumento. Algunosde los instrumentos utilizadospara medir la presi6nsanguineadependen de que el individuopueda palpar (sentir)y auscultar (ofr) los sonidos Korotkoffmientrasque otros instrumentos miden la presi6n sanguinea automaticamente medianteoscilometria,esto es, determinando electr6nicamente el alza y baja de la onda de presi6n sanguinea.Estos nuevos instrumentos electr6nicosque miden la presi6n mediante oscilometriapueden representarpara muchos pacientes una forma convenientey accesiblede proveer seguimientoa su presi6n sanguinea. Estos autores compararon la precisi6nde las lecturasde presi6n sanguineahechas utilizandoun instrumentodigital que detennina la presi6n sanguineadel dedo por metodos oscilometricoscon las lecturas obtenidaspor un esfigmoman6metro. En el estudio se reclutaron76 sujetos,pacientes y personalde un centro de cuidado primario.Se colocaronambos instrumentosde medir presi6n en el brazo Izquierdo y en el dedo de la mano izquierday se tomaron tres lecturasde presi6n sanguineacon cada instrumentoen forma a1tema. EI promedio de todas las lecturasutilizandoel instrumentodigital fue I 19.2/72.4mientrasque con el esfigmoman6metro fue de 118.6/75.3. Las lecturasde presi6n sist61ica no fueron diferentesentre los dos instrumentos(p=0.53), pero las lecturasde presi6n diast61ica resultaronser significativamente diferentes(p=O.OO I). Tambien existierondiferenciasconsiderablesentre las lecturasde presi6n obtenidasen un mismo paciente cuando se usaban los diferentesinstrumentos. AI comparar el oscil6metrodigital
1314 •
Dlep, The Annals ofPharmacotherapy
•
Un appareilde mesure de la tension arteriellepar oscillometrie, dessine pour evaluer la tension arteriellesystemiqueau doigt, a ete compare 11 un sphygmomanometre (SM) 11 colonne de mercure chez 76 sujets patients ou employesd'une c1inique de soins. Apres que les deux appareils eurent ete mis en place, soit sur Ie bras gauche, soit sur un doigt de la main gauche, trois mesures de la tension arterielleont ete obtenues avec chaque appareilen a1temance selon une sequence pre-randomisee. La moyenne de toutes les mesuresetait de I 19.2(72.4 mm Hg pour l'appareil au doigt et I 18.6(75.3 mm Hg pour Ie SM; les lecturesde la tensionsystolique(TS) etaient Iressemblables(p=0.53) mais les lecturesde tension diastolique(TO) etaient differentes(p=O.OOI). On a note des ecarts appreciables entre les lecturesobtenues par les deux appareilschez un merne sujet; les differencesmoyennes (± limitesde tolerance)pour les lecturesobtenues avec l'appareil au doigt comparees au SM etaient de 0.6 ± 33.5 mm Hg pour la TS et -2.9 ± 25.3 mm Hg pour la TO. On a observe aussi une grande variabilite dans les lectures obtenues avec l'appareil au doigt: les differencesentre les minima et maxima dans 39 lecturesde TS (5I pourcent)et 26 lecturesde TO (34 pourcent)etaient de plus de IOmm Hg avec l'appareil au doigt, tandis que 24 lecturesde TS (32 pourcent)et I I lecturesde TO (14 pourcent) differaientelles aussi de plus de IOmm Hg avec Ie SM. En conclusion, la variabilitedans les mesures faites avec l'appareil au doigt etaient significativement differentesde celles prises au sphygmornanometre pour les tensionsdiastoliques(p=0.0048) mais non pour les tensions systoliques(p=0.8729). En se basant sur les resultatsde cette etude, les auteurs ne peuvent recommanderde routine I'emploi d' appareilde mesure de la tension arterielleau doigt.
/991 December, Volume 25
DENYSE DEMERS
